Search

In recent years, battery degradation has become a critical concern in various industries, including electric vehicles, renewable energy systems, and portable electronics. To address this issue, data-driven techniques have emerged as a promising approach for lithium-ion battery (LIB) degradation analysis and estimation. This paper focuses on the application of incremental capacity curves (ICCs) in battery degradation analysis using data-driven techniques. The incremental capacity curve is a powerful tool that provides valuable insights into the capacity degradation of a battery. By analyzing the changes in the ICC over time, it is possible to identify and quantify battery degradation phenomena such as capacity fade, impedance growth, and aging effects. However, manually analyzing ICCs can be time-consuming and subjective, leading to potential errors and inconsistencies. To overcome these challenges, hyperparameter-tuned Long Short-Term Memory (LSTM) techniques are employed to automate the analysis of ICCs and extract meaningful degradation information. These techniques leverage statistical models to process large volumes of ICC data and identify degradation patterns. By training these models on historical data, they can accurately predict battery degradation and estimate the remaining useful life (RUL) of a battery. Further, to enhance the performance of estimation of RUL of the battery. A hyperparameter-tuned LSTM technique has been proposed. The proposed technique has been compared with well-known techniques (i.e. Fully Connected Neural Network (FNN), Artificial Neural Network (ANN), and Convolutional Neural Network (CNN)). The results depict that the proposed robust LSTM technique outperforms well in terms of computational cost and speed. To demonstrate the efficiency of the proposed technique, error analysis has been carried out. The simulation and experimental results depict that the proposed hyperparameter-tuned LSTM model results in very low error indices such as RMSE, MEA, and MAPE as 0.0246, 0.0159 and 1.03 compared with models such as FNN, ANN and CNN. The proposed hyperparameter-tuned LSTM technique depicts a lesser error. By leveraging machine learning and statistical models. The results of this study contribute to the advancement of battery management systems and the optimization of battery usage in various applications.

Data-driven methods have been widely used to estimate the State of health (SOH) of Lithium-Ion batteries (LIBs). However, these methods lack interpretability. In response to this issue, this article proposes a method called Physics-informed neural network (PIFNN) to enhance the interpretability of predictions made by a feedforward neural network (FNN). First, the features are extracted from incremental capacity (IC) curves and differential temperature curves, which can characterize battery aging from different perspectives. Specifically, the peaks of the IC curves (P-IC) reflect the electrochemical reactions that occur during the charge-discharge processes of LIBs. The decline of the P-IC is related to the loss of active materials in LIBs, which is a major cause of the decrease of the SOH. This article converts the monotonous relationship between the P-IC and the SOH into physical constraints to guide the "learning process" of the model. In the prediction process, a physics-constrained secondary "training" is applied to the FNN predictions to further enhance interpretability and improve prediction accuracy. The feasibility of the proposed method is validated using the Oxford and NASA battery datasets. The results indicate that PIFNN effectively improves prediction accuracy and reduces errors to below 1.5 %. • The features are extracted from IC curves and differential temperature curves. • The monotonic relationship between the P-IC and LIBs SOH is transformed into a physical constraint. • A physics-constrained secondary "training" is applied to the FNN predictions. • The feasibility of the proposed method is validated using the Oxford and NASA datasets. [ABSTRACT FROM AUTHOR]